The prognostic value of platelet count related nomogram based on MIMIC-Ⅲ database for critically ill patients
10.3760/cma.j.issn.1671-0282.2022.01.015
- VernacularTitle:基于MIMIC-Ⅲ数据库的血小板计数相关的列线图对重症患者预后的预测价值
- Author:
Lina ZHAO
1
;
Huadong ZHU
;
Yi LI
Author Information
1. 中国医学科学院北京协和医学院，北京协和医院急诊科，复杂重症和罕见病国家重点实验室，北京 100730
- Keywords:
Platelet count;
Critical illness;
Mortality;
Prognosis
- From:
Chinese Journal of Emergency Medicine
2022;31(1):78-84
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To explore the relationship between platelet count (PLT) and organ dysfunction and prognosis in critically ill patients and its prognostic evaluation.Methods:A retrospective cohort study was conducted. The relevant records of 35 860 patients were extracted from the US Intensive Care Database (MIMIC-Ⅲ) from 2001 to 2012. According to the PLT count, patients were divided into the thrombocytopenia group (PLT<100×10 9/L), normal platelet group (100×10 9/L≤PLT≤300×10 9/L), and thrombocytosis group (PLT>300×10 9/L). This study included adult patients between 18 and 89 years old. Patients with survival time less than 24 h and lack of vital signs or PLT values were excluded. The outcome event was the hospital mortality of the patient. Survival was analyzed by the Kaplan-Meier method. The prognostic factors were identified by univariate and multivariate COX analyses. The nomogram to predict hospital mortality was built by the significant prognostic factors. In combination with the important prognostic factors, a nomogram was established to predict the prognosis of critically ill patients in hospital, and the AUC value under the ROC curve was used to assess the discriminative power of the nomogram. Results:Compared with the normal PLT group and the thrombocythemia group, organ dysfunction in the thrombocytopenia group was significantly worse; the SOFA score [3 (2, 5) vs. 2.0 (1, 5) vs. 7 (5, 9)], SAPS-Ⅱ score [31 (23, 41) vs. 32 (23, 42) vs. 38 (30, 50) ], hospital mortality (35.0% vs. 45.2% vs. 54.7%), the incidence of mechanical ventilation (50.3% vs. 41.4% vs. 62.8%), and renal replacement therapy (1.3% vs. 1.3% vs. 6.0%) were significantly higher (all P<0.05). COX regression analysis found that thrombocytopenia was an independent predictor of hospital mortality and was entered into a nomogram after final regressions ( HR=1.477, 95% CI: 1.347-1.691, P<0.01). When the above indicators were brought into the nomogram, the AUC of the nomogram was 0.744. Conclusions:Thrombocytopenia is an independent prognostic predictor of hospital mortality for critically ill patients. PLT-related nomograms have good discrimination, which may help clinicians evaluate the prognosis of hospitalized patients.
